On some problems connected with thermal motion in molecular crystals and a lattice-dynamical interpretation